Web Mining: Analyse fichier log pour demande de credit
Nermine Ali
library(bupaR)## Loading required package: edeaR## Loading required package: eventdataR## Loading required package: processmapR## Loading required package: xesreadR## Warning in library(package, lib.loc = lib.loc, character.only = TRUE,
## logical.return = TRUE, : there is no package called 'xesreadR'## Loading required package: processmonitR## Warning in library(package, lib.loc = lib.loc, character.only = TRUE,
## logical.return = TRUE, : there is no package called 'processmonitR'## Loading required package: petrinetR## Warning in library(package, lib.loc = lib.loc, character.only = TRUE,
## logical.return = TRUE, : there is no package called 'petrinetR'##
## Attaching package: 'bupaR'## The following object is masked from 'package:stats':
##
## filter## The following object is masked from 'package:utils':
##
## timestamplibrary(edeaR)
library(processmapR)
library(eventdataR)
library(readr)
library(tidyverse)## -- Attaching packages ---------------------------------------------------------------------------------- tidyverse 1.2.1 --## v ggplot2 3.1.1 v purrr 0.3.2
## v tibble 2.1.1 v dplyr 0.8.3
## v tidyr 1.0.0 v stringr 1.4.0
## v ggplot2 3.1.1 v forcats 0.4.0## -- Conflicts ------------------------------------------------------------------------------------- tidyverse_conflicts() --
## x dplyr::filter() masks bupaR::filter(), stats::filter()
## x dplyr::lag() masks stats::lag()library(DiagrammeR)
library(ggplot2)
library(stringr)
library(lubridate)##
## Attaching package: 'lubridate'## The following object is masked from 'package:base':
##
## datedata <- read.csv('C:/credit_file.csv', header = T)
summary(data)## Case_ID Activity Resource
## Application_1746793196:28 O_Create Offer :13 User_1 :34
## Application_428409768 :23 O_Created :13 User_19:33
## Application_652823628 :19 O_Sent (mail and online):13 User_3 :17
## Application_1266995739:17 A_Validating :11 User_17:11
## Application_619403287 :17 W_Validate application :11 User_38: 7
## Application_1085880569:16 A_Accepted :10 User_95: 7
## (Other) :49 (Other) :98 (Other):60
## Start_Timestamp Complete_Timestamp
## 2016/01/01 10:51:15.304: 1 2016/01/01 10:51:15.304: 1
## 2016/01/01 10:51:15.352: 1 2016/01/01 10:51:15.352: 1
## 2016/01/01 10:52:36.413: 1 2016/01/01 10:52:36.413: 1
## 2016/01/01 11:16:11.500: 1 2016/01/01 11:16:11.500: 1
## 2016/01/01 11:16:11.549: 1 2016/01/01 11:16:11.549: 1
## 2016/01/01 11:17:31.594: 1 2016/01/01 11:17:31.594: 1
## (Other) :163 (Other) :163
## Variant Variant_index X.case._ApplicationType
## Variant 256:28 Min. : 1.0 New credit:169
## Variant 209:23 1st Qu.: 6.0
## Variant 1 :20 Median : 19.0
## Variant 2 :19 Mean :109.6
## Variant 30 :17 3rd Qu.:209.0
## Variant 6 :17 Max. :388.0
## (Other) :45
## X.case._creditGoal X.case._RequestedAmount Accepted
## Car :53 Min. : 5000 false: 4
## Existing credit takeover:52 1st Qu.: 7000 true : 9
## Home improvement :64 Median :15000 NA's :156
## Mean :13728
## 3rd Qu.:15000
## Max. :35000
##
## Action CreditScore EventID
## Created : 23 Min. : 0.0 Application_1085880569: 1
## Deleted : 12 1st Qu.: 0.0 Application_1266995739: 1
## Obtained : 25 Median : 0.0 Application_1691306052: 1
## statechange:109 Mean : 282.4 Application_1710223761: 1
## 3rd Qu.: 799.0 Application_1746793196: 1
## Max. :1059.0 Application_1878239836: 1
## NA's :156 (Other) :163
## EventOrigin FirstWithdrawalAmount MonthlyCost NumberOfTerms
## Application:75 Min. : 0 Min. :141.4 Min. : 33.00
## Offer :57 1st Qu.: 400 1st Qu.:159.0 1st Qu.: 57.00
## Workflow :37 Median : 3726 Median :200.0 Median : 84.00
## Mean : 8314 Mean :233.1 Mean : 83.54
## 3rd Qu.:15000 3rd Qu.:252.7 3rd Qu.:120.00
## Max. :35000 Max. :498.3 Max. :120.00
## NA's :156 NA's :156 NA's :156
## OfferID OfferedAmount Selected lifecycle.transition
## Offer_117620760 : 4 Min. : 6000 false: 7 complete:144
## Offer_148581083 : 4 1st Qu.:13000 true : 6 start : 25
## Offer_1534843733: 4 Median :15000 NA's :156
## Offer_467024374 : 4 Mean :15108
## Offer_774210695 : 4 3rd Qu.:15700
## (Other) : 24 Max. :35000
## NA's :125 NA's :156
## starttimestamp endtimestamp
## 2016-01-02T11:03:05Z: 3 2016-01-02T11:03:05Z: 4
## 2016-01-02T11:06:53Z: 3 2016-01-02T11:06:53Z: 4
## 2016-01-02T11:30:28Z: 3 2016-01-02T11:43:56Z: 4
## 2016-01-02T11:43:56Z: 3 2016-01-02T11:45:11Z: 4
## 2016-01-02T11:45:11Z: 3 2016-01-02T14:44:54Z: 4
## 2016-01-02T13:38:31Z: 3 2016-01-04T17:23:02Z: 4
## (Other) :151 (Other) :145On peut voir les differentes modalités des variables du fichier log ‘credit file’. Par exemple, les modalités les plus fréquentes du feature Activity sont O_Create Offer, O_Created, O_Sent (mail and online), A_Validating, W_Validate application, A_Accepted et autres.
Conversion des champs date:
Pour aboutir à un fichier event log, nous aurons besoin de convertir les champs starttimesstamp et endtimestamp sous format Datetime.
data$starttimestamp = as.POSIXct(data$`Start_Timestamp`,
format = "%Y/%m/%d %H:%M:%S")data$endtimestamp = as.POSIXct(data$`Complete_Timestamp`, format = "%Y/%m/%d %H:%M:%S")Event log:
Nous avons utilis? deux manières pour la création du fichier event.
Premi?re m?thode:
events <- bupaR::activities_to_eventlog(
head(data, n = 10000),
case_id = 'Case_ID',
activity_id = 'Activity',
resource_id = 'Resource',
timestamps = c('starttimestamp', 'endtimestamp')
)Deuxi?me m?thode:
events <- bupaR::activities_to_eventlog(
data,
case_id = 'Case_ID',
activity_id = 'Activity',
resource_id = 'Resource',
timestamps = c('starttimestamp', 'endtimestamp')
)summary(events)## Number of events: 338
## Number of cases: 10
## Number of traces: 10
## Number of distinct activities: 22
## Average trace length: 33.8
##
## Start eventlog: 2016-01-01 10:51:15
## End eventlog: 2016-02-29 08:00:48## Case_ID Activity Resource
## Length:338 O_Create Offer : 26 User_1 : 68
## Class :character O_Created : 26 User_19: 66
## Mode :character O_Sent (mail and online): 26 User_3 : 34
## A_Validating : 22 User_17: 22
## W_Validate application : 22 User_38: 14
## A_Accepted : 20 User_95: 14
## (Other) :196 (Other):120
## Start_Timestamp Complete_Timestamp
## 2016/01/01 10:51:15.304: 2 2016/01/01 10:51:15.304: 2
## 2016/01/01 10:51:15.352: 2 2016/01/01 10:51:15.352: 2
## 2016/01/01 10:52:36.413: 2 2016/01/01 10:52:36.413: 2
## 2016/01/01 11:16:11.500: 2 2016/01/01 11:16:11.500: 2
## 2016/01/01 11:16:11.549: 2 2016/01/01 11:16:11.549: 2
## 2016/01/01 11:17:31.594: 2 2016/01/01 11:17:31.594: 2
## (Other) :326 (Other) :326
## Variant Variant_index X.case._ApplicationType
## Variant 256:56 Min. : 1.0 New credit:338
## Variant 209:46 1st Qu.: 6.0
## Variant 1 :40 Median : 19.0
## Variant 2 :38 Mean :109.6
## Variant 30 :34 3rd Qu.:209.0
## Variant 6 :34 Max. :388.0
## (Other) :90
## X.case._creditGoal X.case._RequestedAmount Accepted
## Car :106 Min. : 5000 false: 8
## Existing credit takeover:104 1st Qu.: 7000 true : 18
## Home improvement :128 Median :15000 NA's :312
## Mean :13728
## 3rd Qu.:15000
## Max. :35000
##
## Action CreditScore EventID
## Created : 46 Min. : 0.0 Application_1085880569: 2
## Deleted : 24 1st Qu.: 0.0 Application_1266995739: 2
## Obtained : 50 Median : 0.0 Application_1691306052: 2
## statechange:218 Mean : 282.4 Application_1710223761: 2
## 3rd Qu.: 799.0 Application_1746793196: 2
## Max. :1059.0 Application_1878239836: 2
## NA's :312 (Other) :326
## EventOrigin FirstWithdrawalAmount MonthlyCost NumberOfTerms
## Application:150 Min. : 0 Min. :141.4 Min. : 33.00
## Offer :114 1st Qu.: 400 1st Qu.:159.0 1st Qu.: 57.00
## Workflow : 74 Median : 3726 Median :200.0 Median : 84.00
## Mean : 8314 Mean :233.1 Mean : 83.54
## 3rd Qu.:15000 3rd Qu.:252.7 3rd Qu.:120.00
## Max. :35000 Max. :498.3 Max. :120.00
## NA's :312 NA's :312 NA's :312
## OfferID OfferedAmount Selected lifecycle.transition
## Offer_117620760 : 8 Min. : 6000 false: 14 complete:288
## Offer_148581083 : 8 1st Qu.:13000 true : 12 start : 50
## Offer_1534843733: 8 Median :15000 NA's :312
## Offer_467024374 : 8 Mean :15108
## Offer_774210695 : 8 3rd Qu.:15700
## (Other) : 48 Max. :35000
## NA's :250 NA's :312
## activity_instance_id lifecycle_id timestamp
## Length:338 endtimestamp :169 Min. :2016-01-01 10:51:15
## Class :character starttimestamp:169 1st Qu.:2016-01-02 12:05:17
## Mode :character Median :2016-01-02 15:13:23
## Mean :2016-01-08 20:37:05
## 3rd Qu.:2016-01-12 11:00:09
## Max. :2016-02-29 08:00:48
##
## .order
## Min. : 1.00
## 1st Qu.: 85.25
## Median :169.50
## Mean :169.50
## 3rd Qu.:253.75
## Max. :338.00
## Fréquence des évenements:
events %>%
activity_frequency(level = "activity") %>%
plot()
Nous pouvons voir les émails envoyés (O_send) et ceux créé par un offre ou non ont la méme fréquence qui est de l’ordre de 12.5 car dans n’importante quelle cr?ation de compte, tous les utilisateurs vont passer par les mémés étapes c’est à dire phase de creation de la demande d’où sa fréquence élévée. Nous remarquons de plus d’acception, d’inscription et formulaire complet sont les memes alors que ceux du refus sont moins fréquent (refused ou denied).
Traces des activités:
Les fonctions de tracé génériques peuvent être utilisées pour créer des graphiques prédéfinis, pouvant être personnalisés à l’aide de ggplot2.
events %>%
trace_explorer(coverage = 0.9)
events %>%
filter_trace_frequency(perc = 0.9) %>%
process_map()## Warning: Prefixing `UQ()` with the rlang namespace is deprecated as of rlang 0.3.0.
## Please use the non-prefixed form or `!!` instead.
##
## # Bad:
## rlang::expr(mean(rlang::UQ(var) * 100))
##
## # Ok:
## rlang::expr(mean(UQ(var) * 100))
##
## # Good:
## rlang::expr(mean(!!var * 100))
##
## This warning is displayed once per session.events %>%
dotted_chart## Joining, by = "Case_ID"
Le graphique en pointillé personnalisable montre tous les événements par heure. Les relations de priorité entre les activités sont claires: c’est à dire au début on peut voir l’inscription puis le formulaire complet(A_submetted puis Complete application ) Janvier 01 puis les formulaires validés entre le 10 Janvier et fin Janvier puis s’il existe un refus (Denied ) Après le premier Février.
Conclusion:
Nous avons pu analyser un fichier log lié au crédit bancaire des clients afin de pouvoir analyser les activités de ces derniers dans le processus de demande de crédit bancaires. Nous avons utiliser les données présentes comme l’id, type d’activités, le debut et la fin des activités pour analyser les fréquences et les interpréter grace aux outils de visualisation.